Abstract B111: Association of cumulative racialized economic segregation with female breast cancer mortality: A marginal structural model analysis of U.S. counties, 1990–2019

Abstract

Black women experience approximately 40% higher breast cancer mortality than White women, yet the contribution of cumulative, historical structural racism to this disparity remains understudied. Estimating such effects with conventional regression is difficult because of time-varying confounders which are affected by prior exposure. We used marginal structural models to estimate total causal effects of measures of cumulative racialized economic segregation over a 30-year duration on contemporary female breast cancer mortality. In this ecological study of 548 U.S. counties across 48 states and the District of Columbia, the primary exposure was the 30-year measure of cumulative racialized economic segregation, derived using the Index of Concentration at the Extremes (ICE) for Race-Income. The cumulative measure represented each county’s average position in the within-time-point distribution of ICE scores at each of four time-points: 1990, 2000, 2008–2012, and 2015–2019. The outcome was the 2015–2019 age-adjusted female breast cancer mortality rate per 100,000 (ages 50–69) obtained from the CDC WONDER database. To address time-varying confounding, we fit marginal structural models using stabilized inverse-probability-of-treatment weights truncated at the 1st and 99th percentiles. Weights adjusted for historic measures of county poverty, racial composition and population density (logged). The population-weighted outcome model included state-level random intercepts, and confidence intervals were obtained by county-level bootstrap. We further assessed whether adjustment for contemporary healthcare access attenuated the estimated effect. The mean age-adjusted female breast cancer mortality rate was 22.36 deaths per 100,000 (SD = 4.89). After adjusting for time-varying confounders, each 1-standard-deviation increase in cumulative racialized economic privilege was associated with 1.32 fewer deaths per 100,000 (β = −1.320; 95% bootstrap CI, −1.873 to −1.016). Period-specific estimates were consistent across the four time points, with a slight gradient toward the most recent measurement, though the intervals overlapped substantially. A pre-specified sensitivity analysis excluding 21 high-leverage counties with extreme weights yielded a slightly stronger estimate (β = −1.543; 95% CI, −1.753 to −1.314), indicating the effect was not driven by those counties. Adjustment for contemporary healthcare access (primary care and OB/GYN physician density) attenuated the total effect by 16.9%. Counties with a 30-year history of consistent racialized economic privilege had lower breast cancer mortality, and those with consistent disadvantage had higher mortality. Structural interventions (e.g. addressing physician provision) are needed to disrupt the systems that produce entrenched advantage and disadvantage for particular social groups and the resulting mortality disparities. Kamalakanta Gahan, Oyomoare Osazuwa-Peters, Matthew E. Dupre, Hayden B. Bosworth, Hugo Quezada Pinedo, Michael Green, Tomi Akinyemiju. Association of cumulative racialized economic segregation with female breast cancer mortality: A marginal structural model analysis of U.S. counties, 1990–2019 [abstract]. In: Proceedings of the 19th AACR Conference on the Science of Cancer Health Disparities; 2026 Oct 18-21; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2026;35(10 Suppl):Abstract nr B111.